What is our primary use case?
We're using it for big data and storage servers. So mostly Hadoop for big data, Hadoop elastic search, and Ceph storage for our OpenStack private cloud.
The Apollo is performing fairly well. We've run into minor issues, but overall it does the job and we feel it's a good product for the money.
How has it helped my organization?
It's allowed us to benefit from IP-based storage instead of using only fiber channel SAN storage. Also, I don't think we could have afforded that quantity of storage in a SAN array.
What is most valuable?
It's pretty flexible. You can choose how much storage you put on the server. You can have one to three nodes, depending on whether you want more CPU or storage. And we can use the same platform for several use cases: Hadoop, Ceph, and we are considering the server for another use case right now. It's a single solution, we only have to integrate it once and we can use it for several technologies.
What needs improvement?
There should be truly independent nodes for your rack, which can contain three different servers. I like to make sure when a component fails, I don't have to take down all three nodes. This is especially true as we usually have replication between these nodes. It would be a great asset to be able to contain the downtime to one of the nodes.
For how long have I used the solution?
One to three years.
What do I think about the stability of the solution?
It's pretty stable. We've only had very minor issues with it. No major downtime.
The only issues we've really run into so far is that there is a shared battery for all cache controllers in the node. When you have to replace that element, you have to take down all three nodes and not just one. That's something of a design flaw, but it's the only real issue we've had so far.
How is customer service and technical support?
Yes, we've called tech support. Mostly for hardware faults.
What other advice do I have?
When selecting a vendor the most important criteria include
- overall trust in the company
- the financial side, of course, the price of the hardware
- the quality of the support we can expect.
I rate it at eight out of 10. As I said, true independence between the nodes would be an improvement. At least make sure that the nodes aren't dependent on each other. Also, we've had a few difficulties integrating it at first, so I'll stay with an eight.
Test the solution and do a proof of concept until it works with your own integration procedures, the way you install systems, that kind of thing.